I am a big believer in the potential of modern data science techniques to improve real estate investment analysis. Quantitative analysis is already being used by the biggest hedge funds and investment banks and I believe that, as more and more data becomes available, the biggest asset class in the world is ready to benefit from these same techniques.

Sound judgement and expert financial modelling will always remain important (for that you can check my other repo). However, more data driven assumptions are crucial to enhance that analysis.

This repo contains some datasets I've played around in my free time and are solely intended for personal use. A big thanks to the guys at Property Quants who's course taught me many of these techniques and provided some of the datasets. If you are interested in this kind of stuff make sure to check them out!

Currently there are only a few notebooks online. More will follow once I've cleaned up the code a bit. Datasets can be found in the datasets folder.

• avm_singapore.ipynb: Compares ML techniques for building an automated valuation model for real estate in Singapore.
• cl_countries: Compares REIT performance of countries and clusters them in groups.
• cl_london (WIP): Compares London house prices with UK macro economic indicators and clusters the time-periods.
• cl_singapore (WIP): Builds on avm_singapore notebook. Investigates if we can improve our AVM by first clustering the data and building an AVM for each cluster.
• ts_south_africa: Makes forecasts based on time-series data from the South African housing market.